Cryptic Dioecy of Symplocos wikstroemiifolia Hayata (Symplocaceae) and Ternstroemia gymnanthera (Wight & Arn.) Beddome (Pentaphylacaceae) in Taiwan

Androdioecy is an extremely rare breeding system in angiosperms. Theoretical models predicted that the condition for the evolution of androdioecy is stringent and most morphologically androdioecious species were proved to be cryptically dioecious. In Taiwan, Symplocos wikstroemiifolia Hayata and Ternstroemia gymnanthera (Wight & Arn.) Beddome were the only two species documented as morphologically androdioecious. Symplocos wikstroemiifolia had been proposed to be cryptically dioecious, whereas in T. gymnanthera the sexual expression of its breeding system is unclear. This thesis is aimed to validate the existence of androdioecious plants in Taiwan and to increase the knowledge on cryptic dioecy. Thus I laid out experiments of reproductive biology for S. wikstroemiifolia and T. gymnanthera in Taiwan.
In S. wikstroemiifolia, to clarify the functional status of breeding system, patterns of sexual expression were examined in Beichatian Mountain and Pinglin; furthermore, to understand whether S. wikstroemiifolia exhibits a higher level of fruit-set than its hermaphroditic relatives, the natural fruit-set was recorded and compared to that in S. caudata Wall. ex G. Don, a hermaphroditic species co-distributed in Beichatian Mountain. Artificial pollination experiments with emasculation treatments were conducted to evaluate the putative pollinator attraction roles of the stamens or staminodes on morphologically hermaphroditic flowers. Breeding system investigations showed anthers on morphologically hermaphroditic flowers did not produce pollen grains, thus this flower type should be considered as female flowers. Nearly all flowering individuals produced only either male or female flowers, and none of them showed changes of sexes between the two-year field investigations. The sex ratios in the population were slightly toward male-biased, but did not significantly deviate from 1:1. These findings were congruent with the proposed cryptic dioecy status. However, two individuals produced functionally hermaphroditic flowers with limited reproductive abilities. It suggested S. wikstroemiifolia exhibited a nearly complete gender dimorphism but with certain flexibility in sexual expression. In general, there were about 12 stamens per male flowers; and two locules within the ovary of the female flowers, where four ovules (two large, two small) in each locule were found. Male individuals produced racemose cymes and thyrses, whereas female individuals produced racemose cymes only. These morphological characteristics were different from previous discriptions. Patterns of natural fruit-set were not significantly different between S. wikstroemiifolia and S. caudata, which suggested that although only about half of the individuals in the population of S. wikstroemiifolia can produce fruits and seeds, this dioecious species may not maintain a high level of fruit-set as a compensatory strategy. On the other hand, the fruit-set among different artificial pollination treatments also did not differ significantly, which may imply in S. wikstroemiifolia, the staminodes on female flowers might not serve as pollinator attraction.
In T. gymnanthera, because morphologically hermaphroditic flowers produced pollen grains, breeding system investigations largely focused on the pollen fertility from different flower forms. Male flowers showed larger perianth, longer stamens and more pollen grains than morphologically hermaphroditic flowers. Pollen grains from male flowers were tricolporate and prolate to prolate-spheroidal, whereas pollen grains from morphologically hermaphroditic flowers were inaperturate and spheroidal. The mean volume of inaperturate pollen grains was almost nine times as large as that of tricolporate pollen grains. In vitro pollen germination experiments revealed that the inaperturate pollen grains could not germinate in sucrose solution. Artificial pollination experiments showed that in flowers pollinated with pollen grains from morphologically hermaphroditic flowers, only a few of them developed into fruits, and only one seed per fruit were found. In contrast, nearly all flowers pollinated with pollen grains from male flowers developed into fruits containing two to six seeds. These results demonstrated that pollen grains from morphologically hermaphroditic flowers had very high sterility. Morphologically hermaphroditic flowers could be regarded as functionally female flowers. Therefore T. gymnanthera was cryptically dioecious, which exhibited a floral and pollen dimorphism. However, I also observed one male plants producing few functionally hermaphroditic flowers, which may imply an occasional plasticity of sexual expression occur in T. gymnanthera.
In conclusion, there is still no plant species in Taiwan can be proved as functionally androdioecious so far.